In vitro assessment of probiotic attributes for strains contained in commercial formulations

Although probiotics are often indiscriminately prescribed, they are not equal and their effects on the host may profoundly differ. In vitro determination of the attributes of probiotics should be a primary concern and be performed even before clinical studies are designed. In fact, knowledge on the biological properties a microbe possesses is crucial for selecting the most suitable bacteriotherapy for each individual. Herein, nine strains (Bacillus clausii NR, OC, SIN, T, Bacillus coagulans ATCC 7050, Bifidobacterium breve DSM 16604, Limosilactobacillus reuteri DSM 17938, Lacticaseibacillus rhamnosus ATCC 53103, and Saccharomyces boulardii CNCM I-745) declared to be contained in six commercial formulations were tested for their ability to tolerate simulated intestinal conditions, adhere to mucins, and produce β-galactosidase, antioxidant enzymes, riboflavin, and d-lactate. With the exception of B. breve, all microbes survived in simulated intestinal fluid. L. rhamnosus was unable to adhere to mucins and differences in mucin adhesion were evidenced for L. reuteri and S. boulardii depending on oxygen levels. All microorganisms produced antioxidant enzymes, but only B. clausii, B. coagulans, B. breve, and L. reuteri synthesize β-galactosidase. Riboflavin secretion was observed for Bacillus species and L. rhamnosus, while d-lactate production was restricted to L. reuteri and L. rhamnosus. Our findings indicate that the analyzed strains possess different in vitro biological properties, thus highlighting the usefulness of in vitro tests as prelude for clinical research.

Since probiotic properties are often species-or strain-specific 15 , knowledge on the peculiar properties of strains is crucial for choosing probiotic therapies based on specific individual needs.
The present study aimed at evaluating and comparing some in vitro properties of nine microbial strains isolated from commercial formulations sold as containing probiotics. In particular, the ability of these microbes to survive in simulated intestinal fluid, adhere to mucins, produce antioxidant enzymes (i.e. catalase and superoxide dismutase), riboflavin (i.e. vitamin B2), lactase (i.e. β-galactosidase), and d-lactate were investigated. The in vitro analysis of probiotic attributes of microbes isolated from preparations present on the market can be helpful in elucidating the potential beneficial effects provided by these formulations in vivo.

Results
Survival of probiotic microorganisms in simulated intestinal fluid. Bacillus clausii strains NR, OC, SIN, and T, Bacillus coagulans ATCC 7050, Bifidobacterium breve DSM 16604, Limosilactobacillus reuteri DSM 17938, Lacticaseibacillus rhamnosus ATCC 53103, and Saccharomyces boulardii CNCM I-745, declared to be contained in six commercial formulations were isolated and used throughout this study.
The analyzed microbial strains exhibited a very different behavior in simulated intestinal fluid (Fig. 1). The total colony forming units (CFU) number of B. clausii NR and T started to decrease at 4 h (P < 0.01 compared to 2 h) and 2 h (P < 0.001 compared to time 0), respectively. For both strains, the number of CFU recorded after these time points was stable. After 8 h of incubation, a reduction in the total amount of cells compared to the inoculum was registered for both strains. The total CFU number of B. clausii OC decreased after 2 h of incubation in the juice (P < 0.01 compared to 0 min), while started to increase at 6 h (P < 0.01 compared to 4 h). Following an initial decrease at 2 h (P < 0.05 compared to time 0), B. clausii SIN was able to multiply in the fluid starting from 8 h of incubation (P < 0.05 compared to 4 h; Fig. 1) and only a slight reduction of 0.240-Log was observed at this time compared to the inoculum.
A decrease in the total number of B. coagulans and L. reuteri was registered at different times post inoculation (P < 0.01 for B. coagulans at 2 h and 4 h compared to 0 and 2 h, respectively; P < 0.05 for L. reuteri at 4 h compared to 2 h; Fig. 1). Interestingly, the total CFU number of B. breve progressively decreased in the simulated intestinal fluid (decrease of about 1 − Log, P < 0.001 at 2 h compared to 0 min and at 4 h compared to 2 h) and no residual living cells were obtained starting from 4 h of incubation (P < 0.001 at 6 h compared to 4 h). The total number of L. rhamnosus decreased after 2 h and 4 h (P < 0.01 at 2 h compared to 0 min and P < 0.001 at 4 h compared to 2 h). Afterwards, the amount of L. rhamnosus cells was stable over time with a reduction in the total amount of cells of 2.816 − Log at 8 h compared to the inoculum. Lastly, S. boulardii cells persisted in the juice for up to 8 h.
Overall, these results indicate that vegetative cells can behave differently in simulated intestinal fluid, depending on the tested microbial strain. Additionally, our findings highlight the ability of B. clausii NR and T, B. coagulans, L. reuteri, L. rhamnosus, and S. cerevisiae strains to survive in simulated intestinal conditions for up to 480 min. Interestingly, only B. clausii OC and SIN demonstrated the peculiar ability to grow even in the absence of external nutrient sources following an initial decrease in their number. Microbial counts (total CFU number) were performed at the inoculum (0, open bars) and after 2 (diagonally hatched bars), 4 (closed bars), 6 (light grey bars), and 8 (dark grey bars) hours of incubation at 37 °C. Three independent biological replicates with two technical replicates each were performed. Data are expressed as the mean ± standard deviation. For each strain, ANOVA for repeated measures followed by Tukey HSD test was applied to compare the total CFU numbers obtained at each time point. *P < 0.05, **P < 0.01, ***P < 0.001. www.nature.com/scientificreports/ In vitro adhesion of probiotic microbes to mucins. Adhesion of probiotic microorganisms to the gastrointestinal mucus is required for the interaction with host cells, thus allowing probiotics to exert their beneficial activities 6 . Therefore, we investigated the ability of B. clausii, B. coagulans, B. breve, L. reuteri, L. rhamnosus, and S. boulardii strains to adhere to mucins by using a microplate assay 16 . For all B. clausii strains, the CFU/well obtained after incubation on mucins in aerobic conditions was significantly higher than that of the negative controls (P < 0.001 for strain NR; P < 0.01 for B. clausii OC and SIN; P < 0.05 for B. clausii T; Fig. 2a). As shown in Fig. 2b, the B. clausii strains were also able to adhere to mucins following incubation in anaerobic atmosphere (P < 0.01 for B. clausii NR, SIN, and T; P < 0.001 for B. clausii OC compared to negative control). Different behaviors were observed with the other microbial strains. B. coagulans and B. breve were able to adhere to mucins in both aerobic (P < 0.001; Fig. 2a) and anaerobic atmosphere (P < 0.001; Fig. 2b). On the other hand, L. rhamnosus was unable to bind mucins in both conditions, since the total CFU/well recovered from the negative controls was greater than that obtained from mucins (P < 0.01 in aerobic atmosphere and P < 0.001 in anaerobic atmosphere; Fig. 2a,b). While L. reuteri did not adhere to mucins in aerobiosis (Fig. 2a), its adhesion to mucins was substantially increased in anaerobic atmosphere (P < 0.001; Fig. 2b). Conversely, S. boulardii was found to adhere to mucins in aerobic atmosphere (P < 0.01; Fig. 2a) but not in anaerobic conditions (Fig. 2b). and from agar without mucins (i.e. negative control, grey bars) after incubation at 37 °C in anaerobiosis. Three independent biological replicates with two technical replicates each were performed. Data are expressed as the mean ± standard deviation. For each strain, the two-tailed Student's t-test was used to compare the CFU/well obtained on mucins and negative control wells. *P < 0.05, **P < 0.01, ***P < 0.001. www.nature.com/scientificreports/ Production of β-galactosidase by probiotics. Administration of β-galactosidase-producing probiotics can be advantageous in people affected by lactose intolerance for reducing symptoms caused by lactose malabsorption 10 . Therefore, we wondered whether B. clausii, B. coagulans, B. breve, L. reuteri, L. rhamnosus, and S. boulardii could produce β-galactosidase. All B. clausii strains, B. coagulans, B. breve, and L. reuteri strains were able to grow on M63 minimal medium and to degrade X-Gal on M63 and LB agar plates, producing blue colonies (data not shown). In contrast, L. rhamnosus and S. boulardii strains were unable to grow on the minimal medium and formed white colonies on LB agar plates, thus resulting negative for β-galactosidase. The enzymatic activity produced by positive strains was quantified by using the Miller's method 17 . As shown in Fig. 3, the amount of β-galactosidase produced by B. clausii, B. coagulans, B. breve, and L. reuteri strains was significantly higher than that of the negative control (P < 0.01 for B. clausii SIN; P < 0.001 for B. clausii NR, OC, and T, B. coagulans, B. breve, and L. reuteri), thus confirming the ability of these strains to synthetize β-galactosidase. Interestingly, L. reuteri produced the highest amount of enzyme compared to the other β-galactosidase producing strains (P < 0.05 compared to B. breve; P < 0.001 compared to the other strains; Fig. 3). B. coagulans and B. breve synthetized higher enzyme levels than the B. clausii strains (P < 0.001). The enzymatic activity detected for B. clausii SIN was significantly lower compared to the other B. clausii strains (P < 0.05 compared to B. clausii NR and OC; P < 0.001 compared to B. clausii T).

Scientific Reports
Overall, these findings highlight the production of β-galactosidase by the analyzed probiotic B. clausii, B. coagulans, B. breve, and L. reuteri strains.
Production of catalase and superoxide dismutase by probiotic microbes. The production of catalase (CAT) and superoxide dismutase (SOD) by probiotic microorganisms is thought to help the host in reducing oxidative stress 8 . The ability of B. clausii, B. coagulans, B. breve, L. reuteri, L. rhamnosus, and S. boulardii to produce CAT and SOD was evaluated by quantifying these enzymatic activities in cell lysates and culture supernatants collected from actively replicating cells grown in BHIG.
For all the strains, CAT and SOD activities were detected in both cellular compartments. The CAT levels were found to be higher in culture supernatants than in cell lysates (Fig. 4a,b).
In culture supernatants (Fig. 4b), the highest CAT levels were detected for B. clausii OC (P < 0.05 compared to B. clausii SIN, P < 0.01 compared to S. boulardii and B. clausii T, and P < 0.001 compared to B. breve L. reuteri, and L. rhamnosus).
As shown in Fig. 4c,d, the SOD activity was generally higher in cell lysates than in culture supernatants. When SOD was quantified in cell lysates (Fig. 4c), the highest activity was obtained for S. boulardii (P < 0.001 compared to the other strains). By quantifying the SOD activity in culture supernatants (Fig. 4d), the highest SOD activity was measured for B. clausii strains and L. rhamnosus (P < 0.001 compared to the other strains). Significant Figure 3. Quantification of the β-galactosidase activity (Miller Units) produced by probiotic microbes that resulted able to synthesize the enzyme on solid media. P. mirabilis ATCC 12453 was used as negative control in the assay (i.e. negative control). Three independent biological replicates with two technical replicates each were performed. Data are expressed as the mean ± standard deviation. The ANOVA for independent data followed by Tukey HSD test was applied. *P < 0.05, ***P < 0.001. www.nature.com/scientificreports/ differences were found in the extracellular SOD levels among the B. clausii strains. In fact, B. clausii OC showed higher SOD activity compared to NR and T (P < 0.01 and P < 0.05, respectively). Taken together, our results indicate that all the tested strains are sufficiently equipped of antioxidant enzymes that can help in reducing oxidative stress in the host.

Scientific Reports
Production of riboflavin and D-lactate by probiotic microbes. Probiotic microorganisms able to secrete riboflavin are attractive resources that can be useful to compensate host deficiency for this vitamin. Only B. clausii, B. coagulans, and L. rhamnosus strains were found able to produce riboflavin (P < 0.001 compared to the negative control, Table 1). B. coagulans was found to produce the highest level of riboflavin compared to the other strains (P < 0.001). These results highlight the ability of B. clausii NR, OC, SIN, and T, as well as of B. coagulans and L. rhamnosus strains to actively secrete vitamin B2 during growth.
It has been suggested that administration of d-lactate producing strains should be carefully considered in patients at risk of developing d-lactic acidosis 12,14 . . Three independent biological replicates with two technical replicates each were performed. Data are expressed as the mean ± standard deviation. The ANOVA for independent data followed by Tukey HSD test was applied. www.nature.com/scientificreports/ As shown in Fig. 5, only L. reuteri and L. rhamnosus were found able to secrete d-lactate (P < 0.001 compared to the negative control), with L. reuteri producing significantly more d-lactate than L. rhamnosus (P < 0.001).

Discussion
The ability to survive in the presence of bile and adhere to the intestinal mucus are considered key properties that orally administered probiotics should widely possess 6,7 . Alkaline artificial fluids containing bile are commonly adopted as models for evaluating the potential behavior of probiotic microorganisms in the intestinal environment [18][19][20][21][22] . Since the gastrointestinal tract is covered by mucus, mainly formed by mucins 6,23 , incubation of microbes on agar containing mucins in microplates represents a simple and reliable method for investigating their mucin-binding ability and has been adopted in several studies 15,[24][25][26] .
In accordance with its documented bile salt-tolerance and ability to resist to alkaline pH 27 19 . For this reason, bile-salt adapted strains and several microencapsulation techniques have been developed to increase the ability of these bacteria to tolerate harsh conditions and survive in the small intestine 28 .
The finding that B. clausii OC and SIN were able to replicate in the artificial intestinal juice without nutrients lead us to hypothesize that these strains had used the nutrients released by dead cells for multiplication. This behavior agrees with previous studies in which we showed that the spore mixture of strains NR, OC, SIN, and T can germinate and vegetative cells multiply in both the simulated intestinal fluid and human intestine 22,29 . It is important to underline that the tolerance of B. clausii and B. coagulans to simulated intestinal conditions can be improved by administering these strains as spores, which are generally contained in commercial products.
Since oxygen levels progressively decrease along the intestine 30 , we tested the ability of the selected microorganisms to interact with mucins in both aerobic and anaerobic atmosphere. The results obtained for B. clausii NR, OC, SIN, and T and B. coagulans ATCC 7050 could be expected, since genomic analysis of these B. clausii strains and of the B. coagulans S-lac strain revealed the presence of three mucus binding proteins and type-IV filamentous adhesins, respectively 31,32 . In addition, a role of flagellin, the main flagellar filament-component, in mucin adhesion was demonstrated for a probiotic Bacillus cereus strain 33 . In line with studies highlighting the adhesive properties of different B. breve strains 34,35 , B. breve DSM 16604 was found to adhere to mucins. Although the factors involved in B. breve mucin-adhesion are not fully clarified, type IVb tight adherence pili required for glycoprotein interaction have been shown to be essential for the in vivo murine gut colonization 34 .
Among Lactobacillus species, adhesive properties are extremely strain-specific and depend on the presence of specific adhesion factors 6,7,23,[36][37][38][39] . The finding that L. rhamnosus ATCC 53103 was unable to bind mucins correlates with a previous study indicating that some L. rhamnosus strains isolated from different formulations were Figure 5. Production of d-lactate by probiotic microbes. Scatter plot of d-lactate concentration (nmol/ml) values determined in deproteinized culture supernatants of each microbe. Deproteinized BHIG was used as negative control in the assay (i.e. negative control). Three independent biological replicates with two technical replicates each were performed. The mean ± standard deviation was also shown. The ANOVA for independent data followed by Tukey HSD test was applied. ***P < 0.001. www.nature.com/scientificreports/ severely compromised in mucin-binding ability 40 . L. reuteri DSM 17938 and S. boulardii CNCM I-745 adhesion to mucins appeared to be influenced by oxygen levels, in line with their oxygen requirement. Lactose intolerance is a clinical condition caused by β-galactosidase deficiency that results in the inability to degrade lactose contained in milk and derivatives 10,41 .

Scientific Reports
The ability to produce β-galactosidase is widely distributed among the Bifidobacterium and Lactobacillus genera and some strains were shown to bring clinical benefits in people affected by lactose intolerance and malabsorption [41][42][43][44] . However, the production of β-galactosidase by bacteria of these genera appeared strainspecific 43 . The production of this enzyme was also documented for some B. coagulans strains 45,46 .
In line with the evidence that many L. reuteri strains are able to synthesize considerable amounts of β-galactosidase 47,48 , the highest level of this enzyme was evidenced for L. reuteri ATCC DSM 17938. Herein we demonstrate that B. clausii can produce β-galactosidase. Among the tested B. clausii strains, we found that strain SIN produced lower levels of the enzyme. This result agrees with previous studies indicating that the analyzed B. clausii strains exhibit some quantitative differences in their proteome 49,50 . Overall, our results suggest that the tested B. clausii, B. coagulans, B. breve, and L. reuteri strains can potentially be useful in reducing the symptomatology associated with lactose malabsorption, thus providing beneficial effects in individuals affected by lactose intolerance.
Oxidative stress is caused by massive ROS accumulation and is associated with nucleic acid, protein, and lipid modifications, thus leading to cell apoptosis, cellular aging and to the development of several chronic diseases 9,51 . CAT and SOD play a crucial role as antioxidant enzymes, being involved in hydrogen-peroxide and superoxide radical scavenging, respectively 9 . Due to the production of antioxidant enzymes, a role of probiotics in the prevention of oxidative stress and several ROS-linked diseases has been proposed in the last decades 8,9,52 . The finding that the tested Lactobacillus strains showed catalase activity was unexpected, since lactobacilli are generally catalase negative. We can only hypothesize that the adopted assay detects a reduction in the amount of hydrogen peroxide due to the activity of non-catalase enzymes. Other studies will be required to address whether these strains possess an alternative enzymatic activity able to degrade hydrogen peroxide. The results obtained with B. clausii, B. coagulans, B. breve, L. reuteri, L. rhamnosus, and S. boulardii strains indicate that their administration could be helpful in reducing or alleviating ROS accumulation, thus preventing oxidative stress. In particular, the finding that the levels of secreted CAT were higher than those found in the cytoplasm suggests that the analyzed microbes could contribute to counteract extracellular hydrogen peroxide accumulation in vivo, thus exerting a cytoprotective effect.
The B-group and water-soluble vitamin riboflavin is the precursor of flavin adenine mononucleotide and flavin adenine dinucleotide, which act as coenzymes in several redox reactions and are involved in the metabolism of carbohydrates, proteins, lipids, other vitamins, and ketone bodies [53][54][55] . Ariboflavinosis due to dietary inadequacy or gut microbiota dysbiosis can be responsible for a variety of conditions, including cataract, childhood neuropathy, anemia, hypertension, certain types of cancer, and diabetes mellitus 53,54 . For this reason, the use of riboflavin-producing strains and/or administration of probiotic formulations supplemented with exogenous riboflavin can represent a promising source to prevent or ameliorate host ariboflavinosis.
The ability of B. clausii NR, OC, SIN, and T to produce riboflavin confirms a previous study in which these strains were tested for vitamin B2 production using an auxotrophic B. cereus mutant 56 . In accordance with the presence of the riboflavin-metabolic pathway in the B. coagulans genome 57,58 , we found that B. coagulans ATCC 7050 was a very good producer of riboflavin. The detection of riboflavin in L. rhamnosus ATCC 53103 supernatants agrees with previous data reported for this strain 59 . d-lactic acidosis is a neurological disorder typical of patients affected by short bowel syndrome (SBS) and due to massive d-lactate accumulation in the bloodstream 15,60,61 . The administration of d-lactate producing probiotics in the development of d-lactic acidosis has been widely debated and currently appears to be a controversial issue 1,62-65 . However, some authors proposed that d-lactate producing strains should not be administered to patients at risk of developing d-lactic acidosis, including those with SBS and neonates 12,14 . In accordance with the presence of d-lactate dehydrogenase in species belonging to the Lactobacillus genus 15,59,66,67 , we only detected d-lactate in cell culture supernatants of L. reuteri and L. rhamnosus strains.
In conclusion, our findings demonstrate that the investigated probiotic strains are characterized by distinct biological properties in vitro. On the other hand, the only use of in vitro analyses cannot be conclusive to explain the behavior probiotic strains exhibit in vivo, due to the complexity of the gastrointestinal tract. However, we believe that studies investigating the in vitro properties of microbes isolated from commercial products can be of translational relevance, since supporting clinical research in the comprehension of the beneficial effects these formulations may provide in vivo.

Methods
Bacterial strains and culture conditions. Strains used in this study are listed in Table 2. Microbes were isolated from commercial formulations purchased at the local pharmacy. The isolated species were identified by Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry (MALDI-TOF MS) and the strain names as declared by manufactures on the product labels adopted throughout the study. The multi-strain formulation Enterogermina (100 μl) was seeded on Brain Heart Infusion (BHI; Thermo Fisher Scientific, USA) agar plates supplemented with 50 μg/ml rifampicin, 50 μg/ml chloramphenicol, 200 μg/ml streptomycin, or 25 μg/ml tetracycline for selective isolation of B. clausii NR, OC, SIN, and T, respectively. The B. clausii strains used in this study are characterized by a low level of intra-specific genome diversity but showed strain-specific proteomic signaturers 31,48,49 . For mono-strain formulations, microbial isolation was performed as follows. B. coagulans was isolated by seeding 100 μl of Lactò Più on Trypticase Soy agar containing 5% horse blood and plates were incubated at 37   Mucin adhesion in aerobic and anaerobic conditions. Microbial adhesion to mucins was assessed as described by Tsilia et al. 16 with some modifications. Briefly, 100 µl of overnight cultures were inoculated in 25 ml of fresh BHIG and grown to an OD 600 of ~ 1.5. Cultures were centrifuged at 4500 rpm for 10 min at 4 °C and pellets washed two times with sterile PBS. 500 µl of the suspensions were added to 48 well plates (Merck KGaA) containing 600 µl of mucin agar (pH 6.8), constituted by 5% (w/v) mucins from porcine stomach type II (Merck KGaA) and 1% (w/v) bacteriological agar. As negative controls, 500 µl of the same suspensions were inoculated on wells containing 600 µl of 1% (w/v) bacteriological agar. Plates were incubated at 37 °C for 90 min at 50 rpm in aerobic and anaerobic conditions. After incubation, the liquid phase was discarded and wells washed two times with 500 µl of PBS to remove loosely adhered cells. Microbial extraction was performed by mechanical method. The whole solid layers were aseptically transferred in 5 ml of physiological peptone solution and homogenized for 5 min. Aliquots were seeded on agar plates for quantification of adhered bacteria by the plate count method.  www.nature.com/scientificreports/ Quantification of enzymatic activities. The activity of CAT and SOD were quantified in cell lysates and culture supernatants. CAT and SOD activities were measured by using the Catalase Assay kit (abcam, UK) and the Superoxide Dismutase Activity Assay kit (Colorimetric, abcam), respectively, following the manufacturer's instructions.

Preparation of cell lysates and culture supernatants.
Qualitative evaluation of β-galactosidase production was performed by inoculating microbial cells in 5 ml of LB broth containing 0.5 mM IPTG (Thermo Fisher Scientific) for 16 h at 37 °C. Microbial cultures (10 μl) were dropped on M63 synthetic medium (2 g/l (NH 4 ) 2 SO 4 , 13.6 g/l KH 2 PO 4 , 0.5 mg/l FeSO 4 × 7H 2 O, 1 mM MgSO 4 , 0.001 g/l thiamin, 10 g/l lactose, pH 7.0) and LB agar plates, both containing 40 µg/ml X-Gal (Merck KGaA) and 0.5 mM IPTG. Plates were incubated at 37 °C for 48-72 h. Quantification of the β-galactosidase activity was performed as described by Miller 17 . Briefly, microbial cells were inoculated in 5 ml of LB broth supplemented with 0.5 mM IPTG and grown at 37 °C to OD 600 of ~ 0.5. Cells were collected by centrifuging 1 ml of each culture at 8000×g at 4 °C for 5 min and suspended in 1 ml of cold Z buffer (60 mM Na 2 HPO 4 × 7H 2 O, 40 mM NaH 2 PO 4 × H 2 O, 10 mM KCl, 1 mM MgSO 4 × 7H 2 O, and 50 mM β-mercaptoethanol, pH 7.0). Cells permeabilization was performed by adding 20 μl of 0.1% (w/v) sodium dodecyl sulfate (Merck KGaA) and 40 µl of chloroform and by vortexing the tubes for 10 s. Samples (100 µl) were diluted in 900 μl of Z buffer and supplemented with 200 µl of 4 mg/ml orto-nitrofenil-β-galactopyranoside (ONPG, Merck KGaA). The reaction was conducted at 28 °C by monitoring the yellow color development and was stopped by adding 250 µl of 1 M Na 2 CO 3 . β-galactosidase activity was calculated by the equation (Eq. (1)): where OD 420 is the optical density measured at 420 nm, OD 550 the optical density measured at 550 nm, OD 600 the optical density measured at 600 nm, T the reaction time (expressed in min), and V the volume of culture assayed (expressed in ml). Proteus mirabilis ATCC 12453 and Escherichia coli K12 were used as negative and positive controls in the assays, respectively.
Quantification of riboflavin. The amount of riboflavin in culture supernatants was determined by using the Enzyme-linked immunosorbent assay kit for vitamin B2 (Cloud-clone Corp., USA) according to manufacturer's instructions. Sterile BHIG was used as negative control of the assay.
Evaluation of D-lactate production. Before performing d-lactate quantification, 1 ml of culture supernatants was deproteinized by precipitation with 10% (v/v) TCA 6N 68 . The presence of d-lactate in supernatants was evaluated by using the colorimetric d-lactate Assay Kit (abcam) according to manufacturer's instructions. Deproteinized BHIG was used as negative control.
Statistical analysis. For all the experiments, three independent biological replicates with two technical replicates each were performed. Data were expressed as the mean ± standard deviation (S.D). Both statistical analyses and graphs were realized on GraphPad Prism version 8.0.2 (GraphPad Software Inc., USA, https:// www. graph pad. com/ scien tific-softw are/ prism/). For assessing the viability of each microbe in simulated intestinal juice, the one-way analysis of variance (ANOVA) for repeated measure followed by Tukey HSD test was applied to compare the total CFU numbers obtained at each time point. For mucin adhesion experiments, the two-tailed Student's t-test was used to compare the CFU/well obtained for each strain on mucins and negative control wells. For the other assays, ANOVA for independent data followed by Tukey HSD test was used to separately compare the amount of molecules produced by each strain. A two-sided P-value (P) < 0.05 was considered significant.